The nervous system can be thought of as a very complex biological computer whose structural organization has been analyzed for 2,500 years. Yet despite this long history of scientific investigation a great deal of confusion remains about what its basic parts actually are, how they are interconnected, and even what to call them. The long-term objectives of the work proposed here are to develop and implement a mature, robust, user-friendly neuroinformatics workbench on the Web that allows the neuroscience community to access, evaluate, and visualize the neuroanatomical literature. This Brain Architecture Knowledge Management System (BAMS) will facilitate basic research into the cause and treatment of all diseases that affect the brain and mind. A prototype is already available on the Web <http:llbrancusi.usc.edulbkmsl>, and it has four basic modules: brain part Nomenclatures, Relations between parts from different nomenclatures, neuronal Cell Types, and Connections between regions and cell types. The proposed work has four Specific Aims, to enhance or implement the following BAMS components. 1) Nomenclature ontology: a systematic, global ontology for the basic structural divisions of the mammalian nervous system, along with one for the rat central nervous system will be implemented. This neuroanatomy ontology allows comparisons of parts within and between nomenclatures. This work departs from the bias toward a single "standard" brain nomenclature and allows users to determine the meaning of terms used in the literature by particular authors. 2) Homologies: this function allows users to obtain similarity or identity estimates about structures between or within a species, respectively. 3) Networks: this functionality will allow users to view connectional data in schematics, and on flatmaps, atlas levels, and 3-D computer graphics models. 4) Web services protocols: they will allow BAMS to interact with other websites that contain information about neuronal cell types, gene expression patterns, and so on. A project Board of Advisors will help ensure neuroscience community involvement.